I used RVC V2 to make an AI model of a character whose voice is raspy, has a lot of voice cracks and has a very "shouty" voice. I used a dataset which was 70% of voice clips of him yelling and/or talking loudly and the rest were of him talkin…